Domenico Ferrari. Real-time communication in packet-switching wide-area networks. Technical Report TR-89-022, International Computer Science Institute, Berkeley, California, May 1989.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....occur even if bandwidth is not overbooked. For example, two packets with delay bounds of 4 time units and service times of 3 time units each may arrive at a node at the same time. In this situation, it is impossible to meet the delay bounds for both packets. The schedulability test as outlined in [23] is rather complicated. Although connection establishment is a much less frequent event than packet forwarding, and there is usually no real time requirement, it is important to have efficient admission control tests for fast establishment schemes like [82] Compared to the sorted priority queue ....

Domenico Ferrari. Real-time communication in packet-switching wide-area networks. Technical Report TR-89-022, International Computer Science Institute, Berkeley, California, May 1989.

....is predictability, that is, network layer protocols must be able to guarantee the worst case of the communication time between source and destination hosts. The channels of DASH are trying to achieve hard real time communications on the connection oriented communication scheme in the network layer[Ferrari 89] In the case of the channels of DASH, some tests are done and some resources are reserved on each node at connection establishment time to guarantee the worst case of communication time. Even though some resources are reserved on each host, the deadline might be missed because of priority ....

....the channels of DASH, some tests are done and some resources are reserved on each node at connection establishment time to guarantee the worst case of communication time. Even though some resources are reserved on each host, the deadline might be missed because of priority inversion[Tokuda 89] Ferrari 89] claims that it is very difficult to build hard real time communications on the top of connectionless services. Thus, there are too many unspecific factors to achieve the hard real time communications in the large distributed environment. What real time communications mean in large distributed ....

Domenico Ferrari. Real-Time Communication in Packet-Switching Wide-Area Networks. Technical Report TR-89-022, Computer Science Division (EECS), University of California, Berkeley, May 1989.

....the average number of connections in each class using link l at time instant t. 3 Scheduling and Admission Control In Section 2 we assumed that a connection has an end to end delay requirement D, where D is the maximum packet delay. This is called a deterministic delay bound in the literature [10, 11]. Henceforth, instead of the deterministic delay bound, we assume a connection requests an end toend statistical delay bound (D; ffl) i.e. Prob [ packet delay D ] ffl. This is typically required 9 by applications such as voice since they can tolerate some packet loss (a packet is ....

....delay bound, we assume a connection requests an end toend statistical delay bound (D; ffl) i.e. Prob [ packet delay D ] ffl. This is typically required 9 by applications such as voice since they can tolerate some packet loss (a packet is considered lost if its delay exceeds D) [10, 11]. If a connection with an end to end delay requirement (D; ffl) is to be established on an n link route, then we require that each link on the route guarantees ( D n ; ffl n ) 32, 33] which now becomes the local QoS requirement. We now illustrate how to find the schedulability region S of ....

D. Ferrari. Real--Time Communication in Packet--Switching Wide--Area Networks. Technical Report TR-89-022, International Computer Science Institute, Berkeley, California, May 1989.

....minimum. The variation may also be predictable based on a model of how the data is generated. Depending on the equipment used to generate the data, the packet size and rate may be negotiable. Certain applications, such as voice, produce packets at the given rate only for a fraction of the time ([4, 12]) The ON OFF model[13, 14] is used to characterize a data stream. In this model two distinct periods are considered. During the ON period packets arrive deterministically at intervals of T milliseconds. No packets arrive during the OFF period. The ON and OFF periods alternate and are distribute ....

....field contains the requested (utilized or reserved) amount of internal buffer, the fourth field the maximum size of the packets, the fifth field is used to distinguish between a reservation and a confirmed reservation and the last field contains the used output NIF. struct RTCrecord char id[12]; integer bw; integer bmem; integer pktsz; boolean confirmed; char outputNIF; The connection identification (id) is obtained using the source IP address, the destination IP address, the source port and the destination port put back to back one with the other. The obtained value is ensured ....

D. Ferrari, "Real-time communication in packet-switching wide-area networks," Tech. Rep. 89-022, ICSI-International Computer Science Institute, Berkeley, CA, May 1989.

....the same connection. The propagation delay from the source to the destination is a lower bound on the delay a packet would experience. It is This is typically required by applications such as voice since they can tolerate some packet loss (a packet is considered lost if its delay exceeds D s ) [16,17]. For a connection setup request on a multi link route, the requested endto end QoS is divided equally among the links. This is the so called equal allocation policy. For example, if a connection of service s requesting an end to end QoS (D s ; s ) is to be established on an h link route, ....

D. Ferrari. Real--Time Communication in Packet--Switching Wide--Area Networks. Technical Report TR-89-022, International Computer Science Institute, Berkeley, California, May 1989.

....the rate averaging interval might be on the order of several round trip times. The jitter averaging interval would be smaller and would depend on the size of the receive buffer at the destination. Two alternative measures of quality for real time streams have been suggested in recent work [FER], LL] These two alternative measures are the sourceto destination delay bound for a packet and the variance in inter packet gaps observed at a receiver. In transporting real time video or voice, bounds on absolute delay for a packet are not sufficient unless the jitter is also controlled. The ....

D. Ferrari, Real-Time Communication in PacketSwitching Wide-Area Networks, Report No. TR-89022, International Computer Science Institute, Berkeley, May 1989.

....(D s ; s ) where the delay does not include the propagation delay. This QoS requirement is also referred to as packet jitter [16, 50] This is typically required by applications such as voice since they can tolerate some packet loss (a packet is considered lost if its delay exceeds D s ) [15, 16]. If the connection is described by a two state model where it is either in a busy state sending packets back to back at peak rate or in an idle state sending no packets at all, the required bandwidth 2 , denoted by R j c , can be obtained from the following approximation derived in [2, 22, ....

D. Ferrari. Real--Time Communication in Packet--Switching Wide--Area Networks. Technical Report TR-89-022, International Computer Science Institute, Berkeley, California, May 1989.

....of their data generation. The leaky bucket serves to smooth out variations in packet arrivals. For clients who wish to preserve temporal patterns, an approach which tracks transmissions and only intervenes when load specifications are exceeded is more appropriate. Such an approach is described in [Fe89], and this approach is also appropriate for rate control at the nodes. Rate control requires RTIP to retain additional state information for each channel. 2.3. Client Interface A client interface to RCAP could be implemented in the 4.3 BSD UNIX environment as described below. The client ....

D. Ferrari, "Real-Time Communication in Packet-Switching Wide Area Networks, " Rept. No. TR-89-022, International Computer Science Institute, Berkeley, May 1989.

....provided with given local resource allocations at the nodes on the route. Nodes on the route evaluate the end to end delay resulting from per link service requests, and choose the desired level of service. Tenet Suite 2 [4, 13] uses a two pass allocation scheme patterned after Tenet Suite 1 [8, 10], which is extended by providing support for multiparty. Suite 2 uses target sets as receiver abstraction, and provides extensive functionality needed for multiparty, such as sender initiated, and receiver initiated establishment, dynamic membership in target sets, resource sharing among ....

D. Ferrari, "Real-time communication in packet switching wide-area networks," Tech. Rep. TR-89-022, International Computer Science Institute, Berkeley, California, May 1989.

....the propagation 1 This position may be optimal or nearly optimal in terms of availability to client demands, network utilization, or performance criteria. 4 delay in the case of a physical link or the total cumulative delay incurred in the lower level subnetworks in the case of logical link [19]) of each packet has a known and finite bound. It should be mentioned that this assumption may not be satisfied by some links governed by contention based protocols (e.g. Ethernets) however, it can be satisfied on other types of LANs such as token rings (e.g. FDDI rings) Clients communicate ....

Domenico Ferrari. Real-time communication in packet switching wide-area networks. Technical Report TR-89-022, International Computer Science Institute, Berkeley, California, May 1989.

....network s real time clients the desired guarantees by providing real time channels 2 . The basic ideas underlying the scheme were the real time channel paradigm itself, admission control, connection oriented communication, channel rate control, and deadline based scheduling for real time packets [1]. Later, he was joined by another of the authors (Verma) and, in 1989, the Tenet Group was founded with the primary objective of studying real time communication protocols. Shortly thereafter, the remaining authors joined the group. The main activity of the group became, in Fall 1990 and Spring ....

....message is sent, hop by hop, from the source to the destination of the channel. Each RCAP entity maintains a routing table to compute the next hop of channel establishment. At each node along the way where RTIP runs 4 , there is a local RCAP entity, which performs admission control tests [1] [19] If the channel can be supported at this node, the necessary resources are tentatively allocated to the channel and the establish request message (with local parameters appended) is forwarded to the next node. If any node determines that it cannot support the channel s performance ....

[Article contains additional citation context not shown here]

D. Ferrari, "Real-Time Communication in Packet Switching Wide Area Networks," TR-89-022. International Computer Science Institute, Berkeley, May 1989.

....by the Tenet Group, are outlined in Section 5. Section 6 briefly examines related work from the viewpoint of the Tenet principles. Section 7 concludes the paper. 2 The initial Tenet scheme and its extensions 2. 1 The initial scheme The initial Tenet scheme (Scheme 0) is described in detail in [Fer89] FV90a] FV90b] and [VZF91] The network type it assumes has general topology, but consists of identical switches and identical hosts. Thus, Scheme 0 applies to a homogeneous network; internetworking is not supported. The communication abstraction on which Scheme 0 is based is the realtime ....

Domenico Ferrari. Real-time communication in packet switching wide-area networks. Technical Report TR-89-022, International Computer Science Institute, Berkeley, California, May 1989.

....amount of buffer space will have to be statically allocated to each channel to prevent the offending packets from flooding the buffer space of a heavily loaded node and causing packets from other channels to be dropped. A very simple algorithm for distributed rate control is presented in detail in [Ferr89]. One important consequence of this scheme for the algorithms in Section 4 is that we can assume that each channel satisfies its x min and x ave constraints at every node it traverses. 3.3. The channel establishment procedure The channel establishment mechanism we propose for the type of network ....

....saturating the node. If the combination corresponding to all channels through the node being active saturates the node, we will have to consider for the delay bound test a combination in which some of the channels are not active. Details about how such a combination can be arrived at are given in [Ferr89]. For the combination we have selected, let us denote by T the largest of the service times of packets that may traverse the node but do not travel on channels in set U. These include not only those on channels in set V, but also those on other real time channels (e.g. established channels not ....

[Article contains additional citation context not shown here]

D. Ferrari, "Real-Time Communication in Packet Switching Wide-Area Networks ", Tech. Rept. TR-89-022, International Computer Science Institute, Berkeley, May 1989. -20-

....Tenet real time schemes and protocol suites is the real time channel, to be simply called channel in the rest of this paper. In the Tenet Scheme 1, which is the set of algorithms and procedures on which the Tenet Suite 1 [1] is based, a channel is a simplex unicast connection with QoS guarantees [4]. Since guaranteeing performance and reliability bounds requires admission control (and keeping track of how much of each resource is still available for additional channels, an operation usually called somewhat improperly resource reservation ) channels are established prior to their use for ....

.... clients cannot afford long waits for their transmissions to start [5] Scalability: The procedure must work also in very large networks and internetworks [6] Mathematical provability of the guarantees: Real time clients must be offered firm contractual commitments by the service provider [4, 11]. Flexibility in the specification of QoS and traffic parameters: The network layer clients must be given as much freedom as possible in specifying their QoS requirements and traffic descriptions; service classes can be easily defined and easily added on top of a flexible interface, while ....

Domenico Ferrari. Real-time communication in packet switching wide-area networks. Technical Report TR-89-022, International Computer Science Institute, Berkeley, California, May 1989.

....bursty traffic is likely to require real time services. It is therefore useful to investigate the question whether packet switching can be used to provide performance guarantees, and, if so, whether and under what conditions a packet switching solution is more convenient than the other two. In [Ferr89] and [FeVe90] we have provided an answer to the question of feasibility by proposing and evaluating a scheme on which the design of a real time service for an integrated wide area packet switching network could be based. One of the assumptions made in that paper was that the amount of buffer ....

....determining the amount of buffer space that should be allocated to each real time connection in each node on the basis of the characteristics of the traffic on that connection and the bound on the probability of packet loss specified by the client. The algorithm completes the scheme described in [Ferr89] and [FeVe90] and the constructive proof of feasibility that scheme was intended to provide. The important features of the scheme (under the infinite buffer space assumption) are summarized in Section 2. Section 3 presents the buffer space allocation algorithm. The simulation experiments we have ....

[Article contains additional citation context not shown here]

D. Ferrari, "Real-Time Communication in Packet Switching Wide-Area Networks", Tech. Rept. TR-89-022, International Computer Science Institute, Berkeley, May 1989. -

.... to a transfer is not constant, but varies dynamically on demand to accommodate, at least to some extent, peak requests, clients will be interested in adding an average throughput requirement, which should include information about the length of the interval over which the average must be computed [Ferr89a]. Thus, reasonable forms for throughput requirements appear to be the following: i) deterministic throughput bound: q i q min , for all i , where q i is the throughput actually provided by the server, and q min is the lower bound of throughput specified by the client, that is, the minimum ....

D.Ferrari, "Real-Time Communication in Packet-Switching Wide-Area Networks", Tech. Rept. TR-89-022, International Computer Science Institute, Berkeley, May 1989.

....Computer Science Institute. The views and conclusions in this document are those of the authors and should not be interpreted as representing official policies, either expressed or implied, of any of the sponsoring organizations. vide performance guarantees regarding delays or loss rates (see [Ferrari 89] Ferrari 90a] and [Ferrari 90b] for such schemes) it is not very convenient for traffic requiring low delay variation or jitter. A bound on delay jitter is required by both interactive and non interactive applications involving digital continuous media to achieve an acceptable quality of sound ....

....on channel k to be missed. It is possible to determine this worst case situation for each of the existing channels in the node, and to obtain the value of a lower bound on the new channel s delay bound, so that existing delay bounds are not violated. For further details, we refer the reader to [Ferrari 89] The buffer space test verifies that there is sufficient buffer space in the node for the new channel. In general, the buffer space required for the new channel depends on both the local delay bounds and the traffic characteristics of the channel. Since during the forward trip, the delay bound ....

[Article contains additional citation context not shown here]

D. Ferrari, "Real-Time Communication in Packet Switching Wide-Area Networks", Tech. Rept. TR-89-022, International Computer Science Institute, Berkeley, May 1989, submitted for publication.

....x ave constraints declared at the connection establishment time. If the client continues to send data in excess of the promised rate, the number of dropped packets is no longer constrained by the promised loss bound W. A very simple algorithm for distributed rate control is presented in detail in [Ferr89]. One important consequence of this scheme for the algorithms in Section 4 is that we can assume that each channel satisfies its x min and x ave constraints at every node it traverses, at least for the purpose of guaranteeing delay bounds. The buffer space allocation, however, must make sure that ....

....to schedule them so that both meet their deadlines. 7 channels through the node being active saturates the node, we will have to consider for the delay bound test a combination in which some of the channels are not active. Details about how such a combination can be arrived at are given in [Ferr89]. Let the cardinality of this selected combination of channels be K. Let us denote by T the largest of the service times of packets that may traverse the node. With no loss of generality, we assume that the selected channels are numbered 1 through K, according to the order in which their packets ....

[Article contains additional citation context not shown here]

D. Ferrari, "Real-Time Communication in Packet Switching Wide-Area Networks", Tech. Rept. TR-89-022, International Computer Science Institute, Berkeley, May 1989.

....of the field through a systematic study of the complexity and performance of such a scheme. The remainder of this section provides some background for our work. In Section 2, we briefly state some of the principles behind the Tenet Suite; more details, as well as justifications, can be found in [1][2] 3] 4] We then describe the architecture, which arises from a consideration of these principles. Section 3 and Section 4 present the design and the implementation of the individual protocols. In Section 5, we describe our experiments and measurements of the protocols. Section 6 provides a list ....

....real time clients the desired guarantees by providing real time channels 5 . The basic ideas underlying the scheme were the real time channel abstraction itself, admission control, connection oriented communication, channel rate control, and deadline based scheduling for realtime packets [1]. He was later joined by the other authors in designing a protocol suite embodying the Tenet scheme. Several simplifications were made to facilitate the development of the first suite. Most notably, we decided to provide only unicast real time channels, and to introduce multicast capabilities in a ....

[Article contains additional citation context not shown here]

D. Ferrari, "Real-Time Communication in Packet Switching Wide Area Networks," TR-89022. International Computer Science Institute, Berkeley, CA, May 1989.

No context found.

Domenico Ferrari. Real-time communication in packet-switching wide-area networks. Technical Report TR-89-022, International Computer Science Institute, Berkeley, California, May 1989.

No context found.

D. Ferrari "Real-Time Communication in Packet-Switching Wide-Area Networks" Journal of High Speed Networks, Vol. 1, No. 1, 1992, pp. 79-103, 1992. .

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC